← Back to models
DeepSeek logo

DeepSeek

DeepSeek-V4-Pro

Specs

Input price
$1.74/M
Output price
$3.48/M
Context window
1,048.576K tokens

DeepSeek's 1.6-trillion-parameter flagship MoE model - near-frontier intelligence with a 1M-token context window, at a fraction of closed-source cost.

Capabilities

  • 1.6T parameter MoE activating ~49B parameters per token
  • Persistent agentic memory across multi-step tool calls
  • Hybrid Attention (CSA + HCA) cuts KV cache to 10% of standard footprint
  • Manifold-Constrained Hyper-Connections stabilize deep signal propagation

Best for

  • Clinical decision support and case synthesis over long context
  • Repository-wide software engineering and legacy modernization
  • Long-horizon, multi-tool automation agents

Limitations to keep in mind

  • Text-only - cannot natively process medical imaging, audio, or charts
  • Higher latency in Think Max mode (~45-52 tokens/second)
  • 1.6T parameters require immense hardware to self-host

HIPAA-compliant hosting

DeepSeek-V4-Pro is available under our HIPAA-compliant design-partner program, with a signed Business Associate Agreement (BAA), encryption in transit and at rest, and access controls. We are currently onboarding design partners, with general availability coming soon. Run the model behind a unified OpenRouter-style API and swap to another model with a single parameter.

Pricing in context

Open source models like DeepSeek-V4-Pro can be served efficiently on optimized inference infrastructure, with savings passed through to you. Exact savings depend on the model and your volume, but open source inference is typically a fraction of the per-token cost of closed-source frontier models on Azure OpenAI or AWS Bedrock - without cloud egress lock-in or minimum commitments.

Design partner program

Run DeepSeek-V4-Pro under HIPAA

Join the waitlist to be prioritized. We'll reach out with a qualification call and early access.